Discussions

News: Opinion: Is Enterprise Java Open Source real or just a dream?

  1. Due to the development of J2EE Open Source containers we now have full fledge enterprise server containers. Thanks to the service architecture of those containers, we can run other server processes, just directly inside the containers. JOnAS and JBoss for example has a very easy to understand service architecture, so it's quite simple to run your own services inside them.

    The question is whether we have enough server components to be run inside the containers? I imagine that in the near future we only need:

    1) Operating system: a very simple Linux server.
    2) J2EE container: JOnAS or JBoss with all the services needed.
    3) Domain independent services, which can be run inside the
    container: added Open Source Java server components, like:

    - Chat Server: Babylon Server
    - FTP/PSFTP/PSCP Server: http://incubator.apache.org/projects/ftpserver/
    - SSH Server: nothing until now...
    - HTTP Server: Tomcat or Jetty (not Apache, because we want 100% Java solution! This is already done by integrating the web container).
    - Mail/POP/SMTP/IMAP Server: Apache James.
    - CVS Server: nothing until now...
    - LDAP Server: nothing until now...
    - WebDAV Server: nothing until now...
    - Napster Server: http://jnerve.sourceforge.net/
    - DBMS: C-JDBC combined with Java-based DBMS. Could be HSQLDB, McCoi?
    - Other Server Services: ...

    4) Domain dependent components: EJBs, Portlets, Servlets within the container
    - LPortal for Portal and Portlets.
    - OpenUSS for eLearning and community ;-)
    - Compiere for ERP
    - Open For Business for Shop http://www.ofbiz.org/
    - JetTeam, iTracker for software development
    - Ohio for CRM http://www.ohioedge.com/
    - ... all J2EE and Open Source based!

    IMO, this would be very cool stuff. Just run your simple Linux OS, your J2EE container and you will have everything you need to run a real server-side Enterprise Java Open Source System!

    What do you think would this dream become a reality in the near future? Has someone already tried e.g. James instead of sendmail in a huge production system? Or do you think that some of important server components just not reliable enough yet e.g. Java DBMS HSQLDB instead of FireBird ;-)? Or there are some important server components, which are really missing here e.g. CVS server?

    Cheers,
    Lofi Dewanto
    OpenUSS - Open University Support System
    http://www.openuss.org

    Threaded Messages (66)

  2. LDAP Server: LDAPd - http://ldapd.sourceforge.net/
  3. Java LDAP[ Go to top ]

    Looks great! Thanks for the link! I'll watch the project.

    Lofi.
  4. SEDA architecture link[ Go to top ]

    LDAP Server: LDAPd - http://ldapd.sourceforge.net/


    http://www.eecs.harvard.edu/~mdw/proj/seda/ this is the correct link to the SEDA architecuture upon which the ldapd project is based.
  5. Java/J2EE Distribution[ Go to top ]

    Open Source Java/J2EE distribution
    http://www.openenterprisex.org
  6. WebDAV[ Go to top ]

    Slide Advanced WebDAV Servlet
    http://jakarta.apache.org/slide/index.html
  7. FTP and mail inside J2EE container?[ Go to top ]

    Who with sane mind would like to run FTP and mail-server inside J2EE container? What would be the reason behind that? Is is only for acedemic interest?

    In most of the cases FTP, mail and even HTTP servers run on separate machines than J2EE containers for a reason.

    Domain dependent services, I agree, that's what the J2EE containers are for. But, I don't want my application to run slow because my mail server, running inside J2EE server is handling traffic generated from outlook viruses.

    I think, light weight containers is where the future is, avalon family, hivemind etc. It can provide services to ftp server, mail server etc. But, won't force them to co-exist on same container. It's the service centric view, instead of container centric view.

    What do you think?

    Prashant Rane
  8. I think the benefits of cohesion...[ Go to top ]

    as well as the opportunity to enhance the protocols is reason enough to move everything to a container. A container is really just an operating system, anyway.

    I would LOVE to see this revolution become a reality, as I believe it has been the vision from the beginning.

    Best,

    John C. Dale
  9. FTP and mail inside J2EE container?[ Go to top ]

    What if you want to integrate FTP and mail
    into your application? Having them as standalone
    programs means they aren't usable in your application.
    If they were in a container you could chain in
    your own code and possibly override other parts of the
    code.

    For example, i want to do something everytime
    someone ftps a file. I want to limit the number
    of connections. Or i want to prioritize connections
    by customer.
  10. J2EE Container[ Go to top ]

    <quote>
    Who with sane mind would like to run FTP and mail-server inside J2EE container? What would be the reason behind that? Is is only for acedemic interest?
    </quote>

    If you see the architecture of J2EE containers, you can always start any other services inside them. If you don't need the service just turn it off, no big deal. If you only have one maschine you can turn all services, also not a big deal. You have the choice!

    I succesfully run BabylonChat server, James mail server within JOnAS container, not a big deal ;-)

    This is the power of J2EE container: one container for all! And you can turn on or off the services you need or don't need easily!

    Cheers,
    Lofi.
  11. J2EE Container[ Go to top ]

    one container for all!


    this reminds me of quotes from certain famous people.

    - note: not from lord of the rings
  12. I think not[ Go to top ]

    Uh, why not use *just* linux for all those "domain independent services" and concentrate our efforts on the "domain dependent" ones?
  13. Because...[ Go to top ]

    <quote>
    Uh, why not use *just* linux for all those "domain independent services" and concentrate our efforts on the "domain dependent" ones?
    </quote>

    Because:
    - we all love Java ;-) We won't stay here in TSS if we don't really like Java, right?
    - we maybe use another operating systems.
    - we love Linux but don't know how to setup the whole services under Linux.

    Lofi.
  14. You read my mind...[ Go to top ]

    especially about the simple version of linux...just a minimal set of services for translating java requests...hell, though, if you are going to go that far, why not just extend the JVM.

    This, my friends, is the vision, and I hope it becomes available soon.

    Best,

    John C. Dale
  15. You read my mind...[ Go to top ]

    <quote>
    especially about the simple version of linux...just a minimal set of services for translating java requests...hell, though, if you are going to go that far, why not just extend the JVM.
    <quote>

    Cool idea. So at the end we will get a fully Java system! Nice to see someone with the same vision!

    Greets,
    Lofi.
  16. Hallo Lofi...[ Go to top ]

    It is amazing to me that so many developers are blind to the advantages of this.

    Oh well, some will be on the ship, others will drown in the rain.

    At any rate, Vielen Dank for your post. It is refreshing to know I am not alone ;o)

    Java=Elegance=Productivity=Success.

    Best,

    John C. Dale

    "Human history becomes more and more a race between education and catastrophe."
    - H. G. Wells (1866-1946)
  17. Hallo Lofi...[ Go to top ]

    You forgot

    =Slow

    Anyone who bothers to develop GUI apps in Java that are anything more than trivial is doing everyone, themselves included, a disservice.

    My boss hawg machine comes to a grinding halt trying to run Together, etc.

    It's a joke.
  18. Hallo Lofi...[ Go to top ]

    Sartoris: Anyone who bothers to develop GUI apps in Java that are anything more than trivial is doing everyone, themselves included, a disservice. My boss hawg machine comes to a grinding halt trying to run Together, etc.

    Yes, of course. You've been telling us this crap for years. Instead, we should build it in C++. Sure, whatever. While you're smoking that crack pipe, I'm running multiple pure Java GUI applications (and some .NET GUI applications too) and all this on a year-old notebook without problem.

    Anyone who builds a GUI today in anything other than .NET (Windows only) or Java (cross platform) is a friggin' moron.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Clustered JCache for Grid Computing!
  19. Hallo Lofi...[ Go to top ]

    "I'm running multiple pure Java GUI applications (and some .NET GUI applications too) and all this on a year-old notebook without problem."

    Mulitple "Hello Worlds"?

    Sometimes the truth hurts, dude.
  20. Welcome to the present ..[ Go to top ]

    Sartoris: Mulitple "Hello Worlds"? Sometimes the truth hurts, dude.

    Dude, start with multiple pure Java IDEs. Add some Java and .NET utilities and programs. For Java, I run both SWT and Swing apps. Both are more than fast enough on my measly notebook (1Ghz P3). Hmm ... I'm only running one .NET app right this second; it's called SharpReader, and it is probably the most sluggish .NET GUI I've ever seen because it appears to do its disk+network I/O on the GUI thread, or at least stack up the GUI events behind the disk+network I/O operations, so I won't hold it up as a "what you can do". Just like in Java, there are already "pure .NET" dev tools though, built in C# and available via open source, and they perform fine too.

    So I'll stand by my statement: Anyone building a GUI app today should be doing it in Java (if there may be any non-Windows clients) or C#/.NET (if the clients are all Windows and forever shall be.)

    I'm sorry that it is forcing you to learn something newer than MS C/C++ and the MS Windows API, but that's how our industry works, and most everything you know is still applicable (somewhere in the depths) in C#/Winforms. Besides, it's fun to learn new things ... try it, you might actually like it.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Clustered JCache for Grid Computing!
  21. Welcome to the present ..[ Go to top ]

    "So I'll stand by my statement: Anyone building a GUI app today should be doing it in Java "

    My "user experience" is much different than yours it seems. I'll stick by my statement as well.

    "I'm sorry that it is forcing you to learn something newer than MS C/C++ and the MS Windows API, but that's how our industry works, and most everything you know is still applicable (somewhere in the depths) in C#/Winforms. Besides, it's fun to learn new things ... try it, you might actually like it."

    Thanks for the advice. Actually, I have done next to zero C/C++ professionaly in the last almost 3 years. I've been immersed in server side Java writing web applications and being paid handsomely in doing so. Java is quite suitable for web applications and back office stuff.

    I am currently writing a small C# fat client application (and getting paid for it, too!!) so I'm learning something new as you suggest. My impressions so far are good...it is much less tediuos to write it in C# than in Java.

    If I may offer a suggestion to you: never fall in love with a product, technology, language, for it's time to pass will come sooner than you would like.
  22. Hallo Lofi...[ Go to top ]

    Oh well, some will be on the ship, others will drown in the rain.


    So true. Some will refuse to get on the ship, because it wasn't written in Java. :-P
  23. Hallo Lofi...[ Go to top ]

    Zing!

    Personally, I'm trying to figure out why Open Source is such a big deal. I like open source projects and tools and sincerely hope that they'll be around forever. At the same time closed != evil.

    I think as open source projects continue to move forward and be successful, larger entities will express less reluctance to get involved with them. The problem is that open source is often equated with 'free', which frightens corporations because they are in the business of making money. How do you make money off something that is free? Sure, it can be done and has been done, but it breaks a lot of traditional business models.

    I get frustrated by closed systems, but I understand they're a reality. A closed system with a well documented API is often good enough for the task at hand. The benefit of an open source system is that I may not have to implement any cheesy workarounds to get something to work the way I want it. Then again, that's a distinct disadvantage of open source that IT managers are aware of, piddling around in the source is not necessarily desirable for just anyone to be doing.

    The hard fact is, Java may not be around forever. Sure, there are still COBOL programmers out there, but I've watched them migrate away from COBOL since 2000. Java is good for now, but what if something better comes along? I suspect some people will miss it because they're too busy worshipping at the altar of Java to notice.

    I consider myself a practical programmer and I work for a company that is not an IT firm but has a lot of interest in leveraging technology. That means they don't care about Java, they care about results. They care about maintainability to, it's not just a "make it work" mentality. That means a good use of a design pattern will be appreciated, but over-designing everything because "that's the way it has to be done" is not going to win me any accolades.

    I'm not interested in Java or Open Source as a religion. I see the comments some people make in the community and I wonder, do you serve technology or does it serve you?
  24. Open Source = rare support.[ Go to top ]

    I love open source as well, but here is the other reason it's not a catch all. Support is actually poor. Large companies need major issues resolved overnight, and they want vendor accountablity.

    The answer "something is wrong, browse through the apache source, fix it yourself" doesn't work when a ceo has you accountable to solve a problem costing millions over night.

    There are times you WANT accountability, so you can pass that cost onto the close system you bought.
  25. Open Source = rare support.[ Go to top ]

    It's a double edged sword. I suppose it all depends on who is doing your support.

    It's been my experience that if you get good admins, then you don't need the support as much. True, if I have to be the admin for a system, then I don't want to rely on the open source community. After all, I'm not that confident in my admin skills (yet).

    However, the first IT job I worked we had this guy who was a freakin' FreeBSD guru and had spent many years running AIX before that. If he didn't know the answer, it probably couldn't be done. He was the far end of the spectrum though, most of the other admins weren't in his league, but good enough that they knew where to go to ask questions if they were stumped. Some time spent asking question at the right Linux communities was usually enough for them to get back on track.

    My second IT job, the admins were dimestore. The company spent big bucks on IBM hardware, software, and support contracts, but they chintzed on the talent. The problem was, no matter how good the support was (and that's a debatable point right there, 24+ hour response on a Sev-1?) it often relied on our admins to actually implement the solution. It was frightening how often the developers, such as myself, we expected to find out why the servers had crashed.

    Officially, the support may not be there, but get to know your way around and I'll bet you can find people how can help you out. I may not worship at the altar of open source, but I can definitely see some advantages to the communities. Plus, there's no substitute for a good admin.

    Although it is nice to pass along accountability, there's something to be said for being able to come up with a solution without waiting on an outside vendor. Though a lot of companies that distribute open source projects provide support as well.
  26. world domination[ Go to top ]

    Personally, I'm trying to figure out why Open Source is such a big deal.


    <snip>

    > I'm not interested in Java or Open Source as a religion.

    <snip>

    I used to be all about "we will own the space" with JBoss then I calmed down. I focused on building the company JBoss Group to a point where it is a succesful and fast growing business (money in the bank, cash flow positive, stable, record sales back to back bla bla bla). With my bases secured I can be excited about the potential again.

    I am going back to being a bad boy and these days I truly believe the potential of professional open source is complete domination of the sphere you operate in. Already we are unbeatable in OEM and development in JBoss with reported #1 market shares in both, and #3 in deployments and catching up.

    And we are barely getting started, we are fly still... I am assembling top talent in technology and business. In techie guys like remy maucherat, Gavin king and bela ban and with the help of Bob Bickel (the ex GM of Hewlett-Packard's middleware division) we are recruiting some industry ol'timers.

    Seasoned talent is what the company needs to really exploit the potential of professional open source. Bob believes Open Source will revolutionize the industry and so do I.

    The big deal of Open Source, or least "Professional Open Source" like JBoss Group, is that we may prove that the BUSINESS MODEL is viable and dominant, taking the technology further and establishing a model that will be a precedent for future software development (at least in the infrastructure space).

    That ambition is why you and I should care.

    marcf
    Chief Optimist
    JBoss Group LLC
  27. world domination[ Go to top ]

    You should enjoy this.

    The other day during a pre-launch meeting we had a debate over where we were going to deploy the app I was writing. Right now the J2EE environment must be WebSphere or nothing. So there was a concern about licenses, space on existing systems, etc. etc.

    I'm still relatively new to the company, so I jokingly said "So you're saying we can't just throw it on a Linux box running JBoss and call it good?"

    Again, being new I couldn't quite tell if the Technology Manager was annoyed or amused. Maybe he just thought I was naive, but he definitely paused after that one.

    I definitely care about open source, but I don't think all closed systems are bad either. I've been working with Oracle for about 4 years now and I kind of like it. I will admit, if I were in charge I don't know if it would be my first choice, but as a guy who takes what tools he's given, I've had very few problems with it.

    I seriously doubt that open source is going away anytime soon. My computer is riddled with open source software. I know this is true of many developers, and not just in the Java realm.

    At the same time, this attitude of "open source or not at all" that I've encountered seems counter-productive. Especially from line programmers. It would seem more sincere if so many of them didn't change their tone when they get to management. They believe in open source until they're responsible for any mishaps that might occur while using it.

    I see that a pretty standard version of WebSphere runs, what? $8,000? Put that on five servers and you're out $40,000. Pocket change for a large corporation, though a big dent for the little guy. I think the real trick is to convince the big companies to spend that extra money on it's people. You don't even have to spend the entire wad on personnel, but just enough to get good people.

    Unfortunately, that's just crazy talk for a lot of large companies. I guarantee you that the folks at my last job would NEVER go for it. At least where I am at now they are starting to evaluate different server software, and I'm sure app servers will follow close behind.
  28. Well nuts[ Go to top ]

    I'm "Piddlin' Around". That was a test account I set-up when I couldn't remember my password. Didn't realize it was still logged in on this 'puter.

    I was having a case of the Sat'duh'days.
  29. Resin is Open Source.
    OrionServer is Open Source
    Tomcat is Open Source.

    who want's to negotiate a leagal agreement with access to source code to a vendor just in case?

    .V
  30. is really Resin Open Source?[ Go to top ]

    Hi Vic,

    are you sure about Resion being open source? I'm not sure.



    Regards.
    Jesus.
  31. no, Resin is not "open source"[ Go to top ]

    The source is available, but it's not "open source".

    "open source" means "free".
  32. What do you mean?????

    Open Source means FREEDOM but it does NOT mean it free of charge. Writing softwares has a cost.
  33. no, Resin is not "open source"[ Go to top ]

    The source is available, but it's not "open source".

    >
    > "open source" means "free".

    By that definition, is MySQL free, it has run time costs on Windows?

    Open Source means to me that you can get source code free. For Resin, you can.
    Resin is free to evelaute, deveop and for small start ups AFAIK. Of course, Scott would know better. For comercial sites it has a resonable run time fee.
    It is also fast and very popular.

    Open Source does not mean artistic license. They all have a license, it is just resonable.

    .V
  34. Please take a look at http://www.opensource.org/docs/definition.php for a definition of what constitutes an open source product/project.

    As a side note, MySQL has no runtime charges on Windows when used according to the GPL licence.
  35. no, Resin is not "open source"[ Go to top ]

    The source is available, but it's not "open source".

    >
    > "open source" means "free".

    "Free" like in "free radical" or like in "free fall"?
  36. are u sure that OrionServer open source?
    Can you point where i can download source?
  37. OrionServer is Open Source ??[ Go to top ]

    I have not seen any evidence of that.

    Please provide link or information.
    To my knowledge it is quite compact and "cheap" (for a J2EE server), but not an Open Source project (any more)
    Compared to JBoss or JonAS.

    On the other hand, it was (as so many Open Source technologies, and despite al "flames" of Larry Ellison against MySQL and OS in general ;-) embedded into products of commercial companies that completely failed in their J2EE (EJB) efforts before, like Oracle (*bg*)
  38. Java Server Components[ Go to top ]

    There are lots of java server components available and many currently under development (as the other replies have shown). One good place to look is Apache Avalon:

    http://avalon.apache.org

    Avalon started out as the Apache Server Framework but has since expanded into a general IoC component framework. There are many server applications available for Avalon: FTP, Jabber, Mail, HTTP, JMS, ...

    With the recent growth of IoC and component/container frameworks (Avalon, PicoContainer, etc.) I expect any existing void will soon be filled.
  39. so, where is the problem[ Go to top ]

    Things like that, which ones are the barriers that the market must overcome to achieve this ideal world?

    In my opinion:

    - unknowledge, from the managers that all this stuff is available, and how to use it,

    - FEARS!!!!! Consequence usually of the previous one. Who would back my company if things goes wrong?????

    How could a manager take the "risk" of bearing on open source instead paying the compromise of a blue chip company, just in case. Specially managers of largest companies do not mind buy theoretical "safety", despite it is just an illusion.

    For me, that ones are the biggest problems.
  40. Support[ Go to top ]

    <quote>
    How could a manager take the "risk" of bearing on open source instead paying the compromise of a blue chip company, just in case. Specially managers of largest companies do not mind buy theoretical "safety", despite it is just an illusion.
    <quote>

    The support of Open Source J2EE containers is getting better. RedHat joined ObjectWeb (JOnAS) and also recently at the Linux Day in Frankfurt, German SuSe Linux also joined ObjectWeb. It's just a matter of time that the support of Open Source J2EE containers will be outstanding!

    Cheers,
    Lofi.
  41. workflow engine[ Go to top ]

    Look at "Java Business Process Management" for a well designed, open source, j2ee workflow engine : http://jbpm.org

    Regards,
    Tom Baeyens.
    Founder of jBpm.
    Member of the JSR207 expert group "Process Definition for Java".
  42. Dream? No! But what about support?[ Go to top ]

    What do you think would this dream become a reality in the near future?

    Well, if there's companies supporting opensource (Java) projects, then I think this is reasonable to state opensource Java is going to be used for the purpose you mentioned. I think the power behind commercial software is the commercial support organisation (how bad it sometimes may be).

    There's loads of commercial support organisations for Linux. However, for Java opensource software there aren't many (yet). The JBoss Group is giving commercial support like that (the way they do it gives me the creeps, but that doesn't really matter). There are others out there (www.jbosssupport.nl, www.coredevelopers.net), but it's not that big yet.

    The only thing big organisation don't want is to keep maintaining knowledge about all opensource software they're running in house. Consultancies and other support organisation are a must have...

    Alef

    p.s. No, I haven't run any huge enterprise system using James, but there you're giving bad example I think :)... Substitute James with JBoss, Tomcat, Jetty: then yeah, I sure did...
  43. misguided[ Go to top ]

    Java is good. Open source is good. No argument on either point. But why limit yourself to only open source products written in a certain language?

    Apache is a damn good web server. PostgreSQL is a damn good database. They operate along standard protocols (HTTP and SQL, respectively) that make them easy to interact with. Why should I care whether they were written in Java?

    I fail to see the value in reinventing the wheel, just because they didn't use my favorite language.

    -Tim.
  44. misguided[ Go to top ]

    Why should I care whether they were written in Java?


    I do not care what it is writen in. I just use it. pgSQL is great, Linux is great, Sendmail, Apache, Vi, they are all C.

    Of course Tomcat/Struts/iBatis I have to write applcations and maitain them, so... good thing it is writen in my favorite lang.

    .V
  45. Java[ Go to top ]

    <quote>
    Apache is a damn good web server. PostgreSQL is a damn good database. They operate along standard protocols (HTTP and SQL, respectively) that make them easy to interact with. Why should I care whether they were written in Java?
    I fail to see the value in reinventing the wheel, just because they didn't use my favorite language.
    <quote>

    It's simply because you need to rewrite your software sometimes for progress ;-) Why don't we stick to assembler? Why do we need object-orientation? We do we need another IDE (e.g. Eclipse)? This is what we call "progress". The result products would be better than the old ones. I won't call this reinventing the wheel. If the Geronimo project from Apache (J2EE container) tries to build everything from scratch before analyzing another Open Source J2EE containers, than I would call this reinventing the wheel.

    Cheers,
    Lofi.
  46. Java[ Go to top ]

    I seriously doubt that rewriting everything in Java is progress. Certain languages and platforms are better suited for certain functions than others. I love Java, but I wouldn't use it to write a game for anything more powerful than a phone. I also wouldn't bother using it to write anything (like an FTP server) that deals with heavy IO. I could, but the cost associated isn't worth it, and there are already products out there in other languages that work much better. And many of these are also designed well; good design, even good OO design, is not limited to Java. This isn't a matter of saying "let's all go back to writing in assembler." We should write in whatever language and dev platform is best suited to the task at hand, or use other tools that already do what we need better than a potential Java solution.

    Didn't Corel try to create an office suite in Java, and fail miserably? It's not a language well suited for such apps. How many people use the Java Web Server in production deployments compared to Apache?

    Java is a good, clean OO language that was originally designed for applet development and that has excelled in custom, server-side business application development. It has its place, and it does what it does really well (better than any other competitor, in my opinion). But just because I can write nearly anything i want in Java doesn't mean that everything I use _has_ to be written in Java.

    How does the saying go? Once you buy a hammer, everything begins to look like a nail.
  47. Corel[ Go to top ]

    <quote>
    We should write in whatever language and dev platform is best suited to the task at hand, or use other tools that already do what we need better than a potential Java solution.
    <quote>

    I agree with you but IMO, Java is suitable for writing almost *everything*.

    <quote>
    Didn't Corel try to create an office suite in Java, and fail miserably? It's not a language well suited for such apps. How many people use the Java Web Server in production deployments compared to Apache?
    <quote>

    Good question. I wonder, whether they would release the Java Office again. I think if they release it now, it would be different. Java is now a lot more mature compare with 1997 as they tried to release the Java Office.

    Anyway, if you have to write any type of application today (server- and client-side), wouldn't you think about writing it Java?

    Cheers,
    Lofi.
  48. Corel[ Go to top ]

    It depends. As a web app, I certainly would (personally) prefer writing it in Java, but that's only because of my own familiarity with Java. In certain cases, PHP might be a better solution, especially if the app is relatively simple. Or, if there are COM or .NET underpinnings I need to interact natively with, then an ASP.NET solution would probably be better. For thick-client apps, Windows Forms would be a compelling alternative to swing, because of the fact that it's a lot easier to develop with, and performs much better on Windows (or at least doesn't require arcane performance optimizations). For games, I would lean more towards C++, because you need that kind of closeness to the hardware, and the fact that most game devs use C++ (and the fact that there are numerous physics engines or 3d engines out there that are written in C++).

    Java can do (almost) everything, but it's not great at everything. Trying to pretend it is will lead to a lot of flawed solutions.
  49. JBoss Mail + Nukes[ Go to top ]

    JBoss Mail is coming soon. We have a couple of guys contributing towards that. Nukes on JBoss is getting some steam. New stuff for that coming down the pipe too.

    Bill
  50. JBoss Mail + Nukes[ Go to top ]

    Cool, so we are going into the right direction!

    Cheers,
    Lofi.
  51. JBoss Mail + Nukes[ Go to top ]

    JBoss Mail


    Hey, that's nice!!!

    <off-topic>
    About Nukes... I've tried to look into that, but it seems you can only download it including a (tweaked) JBoss distro? No such thing as a download-and-drop-in-EAR file?
    </off-topic>
  52. JBoss Mail + Nukes[ Go to top ]

    Actually we have finished the development fo the forums on nukes. Now we are polishing the set (nukes + modules) to make a downloadable distro.

    julien
  53. I think a better option would be...

    - Embed a JVM into the Linux Kernel (Hey I can always dream)
    - improve the Java API support for the all the current server applications such as
      FTP, Apache and so forth.

    That way we can integrate Java/J2EE into Linux and still use what is already there .
  54. I think a better option would be...

    >
    > - Embed a JVM into the Linux Kernel (Hey I can always dream)
    > - improve the Java API support for the all the current server applications such as
    > FTP, Apache and so forth.
    >
    > That way we can integrate Java/J2EE into Linux and still use what is already there .

    You and Gates have the same general dream, about different things. He wants to embed .NET into Windows Longhorn and he wants to do exactly what you say here. How does it feel to dream the same dreams as Satan himself ? :)
  55. Java/JBoss vs. Longhorn[ Go to top ]

    I think Longhorn is going to be very interesting. It certainly takes the c#/.net version of what we are talking about here with Java/J2EE and pushes it to new heights. MS know that this is going to be a hard sell which is why I think they have started now: possibly 2 years early. Longhorn will either suffer from very slow takeup and possibly fail OR it will be widely accepted and it will make linux etc. look very old fashioned indeed.

    Of course we could all follow like sheep and just use mono to do the same thing under linux OR we could go for an alternative - which this Java/JBoss architecture might be one possible version of.

    The beauty of the open source/community is that we are not betting the farm on any one architecture like MS is - Longhorn will succeed or fail and if it fails MS will have to suffer the consequences (though I doubt it will kill them). The open source world can build lots of different things and natural selection amongst the projects will prevail. The only question is - is this enough to get the invoation needed?

    A Java/JBoss only world is interesting and worth persuing...
  56. Java/JBoss vs. Longhorn[ Go to top ]

    I think Longhorn is going to be very interesting.


    I think that too. And I also think it's gonna rock. Why? Because Microsoft has complete control over the entire stack and will mix the OS with C#/.NET (call it managed code, what ever) in the most efficient and best possible way. How would we do that? Start out a jsr “Integration of the JVM with the OS, whichever might be that” ? And than wait for someone to do it? You know, there are moments when this multi vendor standards based approach is a disadvantage, and I see this as being one of them.

    What do you think?
  57. Java/JBoss vs. Longhorn[ Go to top ]

    A Java/JBoss only world is interesting and worth persuing...


    "plus royaliste que le roi" meaning that you are even more ambitious than we are. Actually that is not entirely true we are very ambitious.

    Just FYI, I am working on the financials of the company for the past 2 days and there are very interesting insights. The company is already very healthy with about 30 employees and cash positive and we never took outside money etc but the future looks rosy. If we execute on the "professional open source" vision we may have a very interesting mid-size company.

    So the discussion of "is it for real" is really not interesting. I mean we run MCI's NOC problem system for the past year, we run the power grid of California, we run playboy.com and the sims (the 2 references I am REALLY proud of) and we ran the Iraq war "real-time information system". We are battle ready. Today it is secure to go with JBoss.

    The real question is how far do we go and how far do we commoditize. I don't like the DELL model, there is no innovation. I want the Professional Open Source model to come up with as much innovation than the companies that made Java. Java was tremendous innovation. We will see the same vitality in Professional Open Source.

    Take the word from me, we are working way ahead of that line (I would characterize this discussion as "late adopters") but we are crossing the chasm, JBoss is secure today, more and more people are finding out.

    WE are about to release FORUMS on JBoss, a BB/Php port on NUKES and JBoss will run it (and hopefully this very forum one day?) and it is an example of another vertical that is commoditized by open source.

    marcf
  58. Application Servers of Mass Destruction[ Go to top ]

    We ran the Iraq war "real-time information system".

    Dude, George Bush is indebted to a Frenchman.
  59. We ran the Iraq war "real-time information system".

    >
    > Dude, George Bush is indebted to a Frenchman.

    Sad. Disgraceful, Shameful for humanity. Anyhow Real time information systems are nice, too bad Bush can´t even operate a 6 digit calculator (I have my doubts about an abacus)
  60. And Open Source[ Go to top ]

    If there was more openness (in politics as well as software development ;) , then the whole war would have not been possible, since the search for weapons of mass destruction would have been a lot easier in an "Open Iraq" by an all "Open USA"

    (like the search for bugs in JBoss or other OS systems, compared to the search of bugs in Windows ;-)
  61. And Open Source[ Go to top ]

    If there was more openness (in politics as well as software development ;) , then the whole war would have not been possible, since the search for weapons of mass destruction would have been a lot easier in an "Open Iraq" by an all "Open USA"

    >
    > (like the search for bugs in JBoss or other OS systems, compared to the search of bugs in Windows ;-)

    You're a f***g genius. Any more of these analogies? Is this how u make decisions?
  62. Look,

    The idea of a JVM embedded into the OS is hardly new and was around long before .Net

    Do you think that Microsoft doesn't take all of the Java communities best ideas and incorporate them into .Net/Windows knowing fully that by the time it get's through the JCP, Microsoft will be way ahead
  63. reinventing wheel[ Go to top ]

    i just find it ridiculous that some people insist that all things must be in java? does this really have a meaning? i dont think so.
    software is not about the language or say means used to develop it
    it is all about design,ideas incorporated,innovations introduced,etc in the product.
    if something is written in lang X and it is solid,proven and has user base
    i will use it no matter which lang i use for own development.
    I dont claim that all languages are equal for developing software in a particular kind of domain.For example I will definitely choose Java over C to develop say new messaging system because of its OO nature,ease of development compared to C,etc.But it will be stupid to try to develop a OS in java
    So the point is that use as much existing,working,proven technology as you can.
    If you have to develop something new, something which will add value to existing stack choose the most appropriate tool which gets the job done whether it is Java or perl or C
    Be pragmatic!
  64. reinventing wheel[ Go to top ]

    <quote>
    Be pragmatic!
    <quote>

    Pragmatic does not mean to become "without vision", right? ;-) Even SAP sees that Java would become the "business" language of SAP, although ABAP/ABAP Object is a very good business language in SAP software. This is all about language and not only programming language. You need a language to communicate. As we all can communicate in English to discuss stuffs like this at TSS. Imagine that you can't understand and write English at all ;-) Wouldn't it be nice to be able to discuss about domain neutral and domain dependent software with "our" language, we understand? It's just easier, yes, a lot more easier!

    Cheers,
    Lofi.
  65. I think one of the reasons why some folks adopt a "Java for everything" mindset is they get hooked on the benefits of the JVM runtime.

    There is something very comforting about knowing that bounds are being checked, memory overwrites are being prevented, memory is being garbage collected, sandboxing is being enforced, etc. Granted, the kernal/CPU will do some of that, but I'm talking intra-user-process.

    Granted, all those features do have a cost in terms of speed, but when you consider the hardware some people are using to read email & browse the web ... ?

    The question is: will "application x" (insert your application of choice there) always be looking for more processing power, or will things reach a point where the benefits of the application-on-VM layer outweigh the cost?

    Ben
  66. One Big Problem = Security[ Go to top ]

    Does anyone ever think about security? Are we supposed to run all these services as root? Are we supposed to start JBoss as root? I don't doubt that having all services run in middleware is the future (we know this), but we have to start thinking about the security of this stuff, or this kind of thing ends up like Windows (Horribly insecure due to fundamental design issues).

    I see 2 major problems in the way of achieving this idea:
    1) Services running on ports below 1024 must start as root. I do not start JBoss or any other Java service as root. I just wont. I run Apache and mod_proxy in front of JBoss (JBoss runs a non-root user). I don't think it's a good idea to run web applications as root. I would be ok with JBoss starting as root if it did what Apache does and just run a small wrapper for the port binding as root and everything else as a non-root user. So maybe JBoss (or other micro-kernels) need a way to start port bindings and services as specific users.

    2) And the problem with that is Java has no way to start threads as different users. The only place I have seen this is in a very old version of the Java Web Server. They actually wrote JNI code to do the user switch. If anyone knows of another way to do this, or if someday it will make it's way into the JDK, please let me know.

    Maybe someone needs to write up a JSR for this. I would be interested in helping.

    -James
  67. One Big Problem = Security[ Go to top ]

    I see 2 major problems in the way of achieving this idea:

    > 1) Services running on ports below 1024 must start as root. I do not start JBoss or any other Java service as root. I just wont. I run Apache and mod_proxy in front of JBoss (JBoss runs a non-root user). I don't think it's a good idea to run web applications as root. I would be ok with JBoss starting as root if it did what Apache does and just run a small wrapper for the port binding as root and everything else as a non-root user. So maybe JBoss (or other micro-kernels) need a way to start port bindings and services as specific users.
    >
    I use this way: java web server accepts connections on 8080 port and all http connections ( 80 port ) are forwarded to
    8080, It must not be a problem for admins to configure network this way.
    I do not think it is more secure to use apache proxies or native wrappers.


    > 2) And the problem with that is Java has no way to start threads as different users. The only place I have seen this is in a very old version of the Java Web Server. They actually wrote JNI code to do the user switch. If anyone knows of another way to do this, or if someday it will make it's way into the JDK, please let me know.
    >
    > Maybe someone needs to write up a JSR for this. I would be interested in helping.
    >
    > -James